Identification of TBK1 inhibitors using a combination of pharmacophore- and machine learning-based approaches

TBK1 is a serine/threonine protein kinase involved in the immune response, metabolism, and tumorigenesis. In this work, we employ a combination of pharmacophore feature-based screening and Machine learning (ML) descriptor-based and other in silico approaches to identify potential TBK1 targeting molecules from the Korean Chemical Database (KCB). The antiproliferative activity of the top selected compounds was tested in a human breast cancer cell line (MDA-MB-231). Seven molecules were identified that displayed significant IC50 values (below 5 μM). These molecules can be used as a starting point to further explore, design and optimize novel inhibitors targeting TBK1 protein in cancer therapeutics. Moreover, our study can be used as a benchmark to identify novel small molecules targeting TBK1 protein from different free and commercial chemical databases.